Full 3-D electromagnetic subsurface imaging using ground penetrating radar

Abstract

This work is concerned with the 3-D imaging problem of buried objects via enhanced microwave tomography ground penetrating radar (GPR) surveys. In particular, we propose a model based imaging approach, which exploits the Born approximation and accounts for the vectorial nature of the scattering phenomenon as well as for the presence of the air-soil interface. Moreover, a truncated singular value decomposition (TSVD) inversion scheme is applied to achieve stable and accurate results. The advantages offered by a full 3-D inversion algorithm with respect to the commonly adopted strategy, which produces 3-D images by interpolating 2-D reconstructions, are assessed against experimental data gathered in laboratory controlled conditions.

A comparison between two measurement configurations for full 3D GPR imaging

“…Nowadays, owing to the improvement of the computational resources, full 3D inversion approaches accounting for the inherent vector nature of the scattering phenomenon are becoming of huge interest. These methods are, indeed, able to offer improved performance in terms of image focusing both in controlled and on-field experiments [8], [9].…”

Fast 3D subsurface imaging with stepped-frequency GPR

“…There have been several twodimensional (2D) imaging analyses 1,2 , as well as several three-dimensional (3D) analyses 3,4 . The 3D imaging problem is typically very difficult because of the high computational burden imposed by a fully 3D inversion, but significant progress has been made on this front by exploiting the special structure of the forward operators when the data is collected in a uniform fashion 5 .…”